Yiqing Xu

National University of Singapore. yiqing[dot]xu[at]u[dot]nus[dot]edu

prof_pic.jpg

NUS School of Computing, COM1, 13, Computing Dr

Singapore 117417

I’m Yiqing Xu, a CS Ph.D. candidate at National University of Singapore advised by Prof. David Hsu. Previously, I obtained double degrees in Computer Science and Applied Mathematics from National University of Singapore.

I am currently a visiting Ph.D. student at MIT CSAIL, advised by Prof. Leslie Kaelbling and Prof. Tomás Lozano-Pérez.

research highlights

My research on Robot Learning focuses on translating human objectives into signals for machine optimization. I design algorithms and learn representations to align AI agents more closely with human objectives, enabling them to better understand and assist humans.

I’ve been working on algorithm research and theoretical analysis of inferring reward functions (Inverse RL) from limited data. Recently, I’ve used large language models to learn reward functions for motor skills through self-alignment. My ongoing work addresses the challenge of defining and solving robotics tasks with ambiguous and multi-factorial objectives, such a ``tidying up a table”, using more flexible and comprehensive frameworks than conventional reward mechanisms.

selected publications

  1. Receding Horizon Inverse Reinforcement Learning
    Yiqing Xu, Wei Gao, and David Hsu
    Advances in Neural Information Processing Systems, 2022
    arXiv
  2. Learning Reward for Physical Skills using Large Language Model
    Yuwei Zeng, and Yiqing Xu
    Conference on Robot Learning, LangRob workshop, 2023
    arXiv
  3. "Tidy Up the Table": Grounding Common-sense Objective for Tabletop Object Rearrangement
    Yiqing Xu, and David Hsu
    2023
    arXiv
  4. On the Effective Horizon of Inverse Reinforcement Learning
    Yiqing Xu, Finale Doshi-Velez, and David Hsu
    In submission, 2023
    arXiv